# Lockheed: Data Exploration and Visualization
# Author: Hao Zhong
# Install packages (commented out)
# install.packages("ggplot2")
# install.packages("GGally")
# Set working directory to wherever this file locates
library(rstudioapi)
# setwd(dirname(getActiveDocumentContext()$path))
getwd()
## [1] "/Users/Hao/Projects/RPI-DARPA-DS"
# Use with caution as this will automatically clean up the whole workspace and
# redo all the data import and preparation steps. If you feel unsure, please run
# this file "Lockheed.R" manually and suit to your purpose.
source("Lockheed.R")
## Warning: NAs introduced by coercion
################################################################################
library(ggplot2)
# Add `Gic.Level` to `mydata` for convenience in using `ggplot2::facet_wrap`
mydata <- cbind(mydata, Gic.Level)
#
# 1.1 Barplot: One Categorical Variable (`ggplot2::geom_bar`)
#
# Define the categorical variable by `aes(x = ...)`
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation))
ggplot(mydata) + geom_bar(aes(x = Contaminate.Type))
# Use `aes(fill = ...)` to color each bar by its own category
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Surface.Preparation))
ggplot(mydata) + geom_bar(aes(x = Contaminate.Type, fill = Contaminate.Type))
# Use `ggplot2::facet_wrap` to facet into multiple plots by another factor
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Surface.Preparation)) + facet_wrap(~ Contaminate.Type)
ggplot(mydata) + geom_bar(aes(x = Contaminate.Type, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation)
#
# 1.2 Stacked Barplot: Categorical vs Categorical (`geom_bar` w/ aes(fill = ...))
#
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Contaminate.Type))
ggplot(mydata) + geom_bar(aes(x = Contaminate.Type, fill = Surface.Preparation))
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Gic.Level))
# Facetting by another factor (`facet_wrap`)
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Contaminate.Type)) + facet_wrap(~ Contaminate.Type, ncol = 2)
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Contaminate.Type)) + facet_wrap(~ Gic.Level, ncol = 1)
ggplot(mydata) + geom_bar(aes(x = Contaminate.Type, fill = Surface.Preparation)) + facet_wrap(~ Surface.Preparation, ncol = 1)
ggplot(mydata) + geom_bar(aes(x = Contaminate.Type, fill = Surface.Preparation)) + facet_wrap(~ Gic.Level, ncol = 1)
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Gic.Level)) + facet_wrap(~ Contaminate.Type)
ggplot(mydata) + geom_bar(aes(x = Surface.Preparation, fill = Gic.Level)) + facet_wrap(~ Gic.Level)
# Note that `ncol = k` sets the facet grids in k column(s).
# For other usages/ please go to `?facet_wrap`
#
# 1.3 Boxplot: Categorical vs Numerical (`ggplot2::geom_boxplot`)
#
# `fill = NULL`
ggplot(mydata) + geom_boxplot(aes(x = Gic.Level, y = Adhesive.Out.Time, fill = NULL))
ggplot(mydata) + geom_boxplot(aes(x = Gic.Level, y = Prep..to.Bond.Time, fill = NULL))
ggplot(mydata) + geom_boxplot(aes(x = Gic.Level, y = Contamination.Amount, fill = NULL))
# `fill` same as `x`
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Adhesive.Out.Time, fill = Contaminate.Type))
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Prep..to.Bond.Time, fill = Contaminate.Type))
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Contamination.Amount, fill = Contaminate.Type))
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Gic, fill = Contaminate.Type))
# Consider facetting using another categorical variable different from `x`
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Adhesive.Out.Time, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Prep..to.Bond.Time, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Contamination.Amount, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
ggplot(mydata) + geom_boxplot(aes(x = Contaminate.Type, y = Gic, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
# `fill` different from `x`
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Adhesive.Out.Time, fill = Gic.Level))
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Prep..to.Bond.Time, fill = Gic.Level))
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Contamination.Amount, fill = Gic.Level))
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Gic, fill = Gic.Level))
# Consider facetting using another categorical variable different from `x`
# In this case it avoids a overly dense plot
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Adhesive.Out.Time, fill = Gic.Level)) + facet_wrap(~ Gic.Level, ncol = 1)
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Prep..to.Bond.Time, fill = Gic.Level)) + facet_wrap(~ Gic.Level, ncol = 1)
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Contamination.Amount, fill = Gic.Level)) + facet_wrap(~ Gic.Level, ncol = 1)
ggplot(mydata) + geom_boxplot(aes(x = Surface.Preparation, y = Gic, fill = Gic.Level)) + facet_wrap(~ Gic.Level, ncol = 1)
#
# 1.3.* Violin Plot: Categorical vs Numerical (`geom_violin`)
#
# Violin plot (using `ggplot2::geom_violin`) can be seen as a "smoothed" version
# of box plot, however not recommended in this case. Its syntax is identical to
# `geom_boxplot` therefore you can easily change from a boxplot to a violin plot.
#
# Examples:
ggplot(mydata) + geom_violin(aes(x = Surface.Preparation, y = Adhesive.Out.Time, fill = Surface.Preparation))
ggplot(mydata) + geom_violin(aes(x = Surface.Preparation, y = Prep..to.Bond.Time, fill = Surface.Preparation))
ggplot(mydata) + geom_violin(aes(x = Surface.Preparation, y = Contamination.Amount, fill = Surface.Preparation))
ggplot(mydata) + geom_violin(aes(x = Surface.Preparation, y = Gic, fill = Surface.Preparation))
#
# 2.1 Histogram: One Numerical Variable (`ggplot2::geom_histogram`)
#
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# Use `facet_wrap` to facet into multiple plots by another factor
# Facet by Contaminate.Type
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time)) + facet_wrap(~ Contaminate.Type, ncol = 2)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time)) + facet_wrap(~ Contaminate.Type, ncol = 2)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount)) + facet_wrap(~ Contaminate.Type, ncol = 2)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic)) + facet_wrap(~ Contaminate.Type, ncol = 2)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# Facet by Surface.Preparation
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# Facet by Gic.Level
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#
# 2.1.* Density Plot: One Numerical Variable (`geom_density`)
#
# Basically a smoothed version of histogram, just as violin plot is to boxplot.
# Examples:
ggplot(mydata) + geom_density(aes(x = Contamination.Amount))
ggplot(mydata) + geom_density(aes(x = Gic))
ggplot(mydata) + geom_density(aes(x = Contamination.Amount)) + facet_wrap(~ Surface.Preparation, ncol = 1)
ggplot(mydata) + geom_density(aes(x = Gic)) + facet_wrap(~ Surface.Preparation, ncol = 1)
ggplot(mydata) + geom_density(aes(x = Contamination.Amount)) + facet_wrap(~ Gic.Level, ncol = 1)
#
# 2.2 Stacked Histogram: Numerical vs Categorical (`geom_histogram` + `aes(fill = ...)`)
#
# fill = Surface.Preparation
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Surface.Preparation))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Surface.Preparation))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Surface.Preparation))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Surface.Preparation))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Contaminate.Type
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Contaminate.Type))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Contaminate.Type))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Contaminate.Type))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Contaminate.Type))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Gic.Level
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Gic.Level))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Gic.Level))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Gic.Level))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Gic.Level))
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# Facetting by another factor (`facet_wrap`)
# fill = Surface.Preparation, facet by Contaminate.Type
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Surface.Preparation)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Surface.Preparation)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Surface.Preparation)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Surface.Preparation)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Surface.Preparation, facet by Gic.Level
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Surface.Preparation)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Surface.Preparation)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Surface.Preparation)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Surface.Preparation)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Contaminate.Type, facet by Surface.Preparation
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Contaminate.Type)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Contaminate.Type, facet by Gic.Level
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Contaminate.Type)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Contaminate.Type)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Contaminate.Type)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Contaminate.Type)) + facet_wrap(~ Gic.Level, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Gic.Level, facet by Contaminate.Type
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Gic.Level)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Gic.Level)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Gic.Level)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Gic.Level)) + facet_wrap(~ Contaminate.Type, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# fill = Gic.Level, facet by Surface.Preparation
ggplot(mydata) + geom_histogram(aes(x = Adhesive.Out.Time, fill = Gic.Level)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Prep..to.Bond.Time, fill = Gic.Level)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Contamination.Amount, fill = Gic.Level)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(mydata) + geom_histogram(aes(x = Gic, fill = Gic.Level)) + facet_wrap(~ Surface.Preparation, ncol = 1)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#
# 2.2.* Overlapping Density Plots: Numerical vs Categorical (`geom_density` + `aes(fill = ...)`)
#
# fill = Surface.Preparation
ggplot(mydata) + geom_density(aes(x = Adhesive.Out.Time, fill = Surface.Preparation), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Prep..to.Bond.Time, fill = Surface.Preparation), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Contamination.Amount, fill = Surface.Preparation), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Gic, fill = Surface.Preparation), alpha = 0.5)
# fill = Contaminate.Type
ggplot(mydata) + geom_density(aes(x = Adhesive.Out.Time, fill = Contaminate.Type), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Prep..to.Bond.Time, fill = Contaminate.Type), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Contamination.Amount, fill = Contaminate.Type), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Gic, fill = Contaminate.Type), alpha = 0.5)
# fill = Gic.Level
ggplot(mydata) + geom_density(aes(x = Adhesive.Out.Time, fill = Gic.Level), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Prep..to.Bond.Time, fill = Gic.Level), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Contamination.Amount, fill = Gic.Level), alpha = 0.5)
ggplot(mydata) + geom_density(aes(x = Gic, fill = Gic.Level), alpha = 0.5)
#
# 2.3 Scatterplot: Numerical vs Numerical (`geom_point`)
#
# Exploring Gic vs Contamination.Amount
# Basic scatterplot: Gic vs Contamination.Amount
ggplot(mydata) + geom_point(aes(x = Gic, y = Contamination.Amount))
# Color points by a factor
ggplot(mydata) + geom_point(aes(x = Gic, y = Contamination.Amount, col = Contaminate.Type))
ggplot(mydata) + geom_point(aes(x = Gic, y = Contamination.Amount, col = Surface.Preparation))
# Separate into multiple facets by a factor
ggplot(mydata) + geom_point(aes(x = Gic, y = Contamination.Amount, col = Contaminate.Type)) + facet_wrap(~ Surface.Preparation)
ggplot(mydata) + geom_point(aes(x = Gic, y = Contamination.Amount, col = Surface.Preparation)) + facet_wrap(~ Contaminate.Type)
# For Adhesive.Out.Time and Prep..to.Bond.Time, the values are widely apart and
# thus awkard to visualize straighforward
ggplot(mydata) + geom_point(aes(x = Adhesive.Out.Time, y = Prep..to.Bond.Time))
# If we try to separate them into different facets and use free axis scales by
# using `facet_wrap(..., scales = "free")`, it is revealed that within each or
# each combination of categories, these 2 times have a correlating pattern.
ggplot(mydata) + geom_point(aes(x = Adhesive.Out.Time, y = Prep..to.Bond.Time)) + facet_wrap(~ Surface.Preparation, scales = "free")
ggplot(mydata) + geom_point(aes(x = Adhesive.Out.Time, y = Prep..to.Bond.Time)) + facet_wrap(~ Contaminate.Type, scales = "free")
ggplot(mydata) + geom_point(aes(x = Adhesive.Out.Time, y = Prep..to.Bond.Time)) + facet_wrap(Surface.Preparation ~ Contaminate.Type, scales = "free")
#
# 2.3.* 2-D Density Plot: Numerical vs Numerical (`geom_density_2d`)
#
ggplot(mydata) + geom_density_2d(aes(x = Gic, y = Contamination.Amount))
ggplot(mydata) + geom_density_2d(aes(x = Gic, y = Contamination.Amount)) + facet_wrap(~ Surface.Preparation)
ggplot(mydata) + geom_density_2d(aes(x = Gic, y = Contamination.Amount)) + facet_wrap(~ Contaminate.Type)
ggplot(mydata) + geom_density_2d(aes(x = Adhesive.Out.Time, y = Prep..to.Bond.Time)) + facet_wrap(Surface.Preparation ~ Contaminate.Type, scales = "free")
## Warning: Computation failed in `stat_density2d()`:
## bandwidths must be strictly positive
## Warning: Computation failed in `stat_density2d()`:
## bandwidths must be strictly positive
################################################################################
# What we have done above:
# 1) Barplot on one categorical variable
# 2) Represent how each bar is made up by another variable
# - second variable is factor: add fill color in barplot
# - second variable is numerical: as the y-axis in violin plot
# 3) Make multiple facets by a 3rd (even a 4th, if it makes sense yet still tidy) variable (factor)
# Comment:
# From domain expert, also inspired by our intuition from the categorical variables
# on experiment design, it appears reasonable that we perform some data analytics
# later for each combination of Surface.Preparation and Contaminae.Type, especially
# those analysis that are more for numerical variables.
################################################################################
################################################################################
#
# Contigency Table: Factor1 vs Factor2 (Using `table()`)
#
# Build a contingency table of the counts at each combination of factor levels
table(mydata$Contaminate.Type, mydata$Surface.Preparation)
##
## FPNG N3 P & S Pla
## Fre 40 40 42 41
## GL 11 12 11 9
## GPA 12 12 12 12
## GT 11 12 6 12
## Law M P 18 12 0 0
## Min 12 12 12 12
## Na T 12 12 12 12
## Oct 11 12 12 12
## PBJ 7 12 12 9
table(Gic.Level, mydata$Surface.Preparation)
##
## Gic.Level FPNG N3 P & S Pla
## High 58 59 57 51
## Low 76 77 62 68
table(mydata$Contaminate.Type, Gic.Level)
## Gic.Level
## High Low
## Fre 70 93
## GL 21 22
## GPA 19 29
## GT 17 24
## Law M P 14 16
## Min 23 25
## Na T 23 25
## Oct 25 22
## PBJ 13 27
################################################################################
#
# Pairwise scatterplots
#
# Use `plot()`
plot(mydata[c(doe.all, "Gic")])
# Use `GGally:ggpairs``
library(GGally)
# References:
# http://dkhramov.dp.ua/Comp.PlotMultidimensionalData#.WWepucbMyu4
# https://www.rdocumentation.org/packages/GGally/versions/1.3.1/topics/ggpairs#Details
# "400000" is a cutoff Adhesive.Out.Time value to subset the fake data
ggpairs(data = mydata[which(mydata$Adhesive.Out.Time < 400000), ],
columns = c(doe.all, "Gic"),
upper = list(
continuous = "cor",
combo = "facethist",
discrete = "facetbar"
),
lower = list(
continuous = "points",
combo = "box",
discrete = "ratio"
),
diag = list(
continuous = "barDiag"
),
mapping = aes(color = Contaminate.Type)
)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# Include only 1 categorical variable for clarity: Contaminate.Type
ggpairs(data = mydata[which(mydata$Adhesive.Out.Time < 400000), c(doe.all, "Gic")],
columns = c(doe.num, "Contaminate.Type", "Gic"),
upper = list(
continuous = "cor",
combo = "facethist",
discrete = "facetbar"
),
lower = list(
continuous = "points",
combo = "box",
discrete = "ratio"
),
diag = list(
continuous = "barDiag"
),
mapping = aes(color = Contaminate.Type)
)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# Include only 1 categorical variable for clarity: Surface.Preparation
ggpairs(data = mydata[which(mydata$Adhesive.Out.Time < 400000), ],
columns = c(doe.num, "Surface.Preparation", "Gic"),
title = "Pairwise plots of DoE variables and Gic (Using GGally:ggpairs)",
upper = list(
continuous = "cor",
combo = "facethist",
discrete = "facetbar"
),
lower = list(
continuous = "points",
combo = "box",
discrete = "ratio"
),
diag = list(
continuous = "barDiag"
),
mapping = aes(color = Surface.Preparation)
)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# There are other possible options include `pairs`, `lattice::splom`.
################################################################################
# Next start from here ...
################################################################################
#
# Radial plot
#
# library(plotrix)
# radial.plot(scale(mydata[, c(doe.num, "Gic")]), # данные
# labels=names(mydata[, c(doe.num, "Gic")]), # подписи к осям
# rp.type="p", # тип линии ("p" - полигон)
# lwd="1", # толщина линии
# line.col=mydata[, 4], # цвет линии (номер цвета в палитре)
# rad.col="lightblue" # цвет осей
# )
################################################################################
#
# Andrews plot
#
# install.packages("andrews")
# library(andrews)
# andrews(mydata, clr = 4)
# andrews(mydata, type = 4, clr = 5)
################################################################################
#
# Parallel coordinates plot
#
# Use GGally::ggparcoord
# mydata.scaled <- cbind(scale(mydata[, doe.num]), mydata[, c(doe.cat, "Gic")])
# ggparcoord(mydata.scaled, columns = c(1,2,3,6), groupColumn = "Contaminate.Type", shadeBox = "red",
# scale = 'globalminmax', boxplot = TRUE, alphaLines = .5) +
# facet_wrap(~ Surface.Preparation)
# Use parcoords::parcoords
# http://www.buildingwidgets.com/blog/2015/1/30/week-04-interactive-parallel-coordinates-1
# devtools::install_github("timelyportfolio/parcoords")
# library(parcoords)
# parcoords(mydata, rownames = FALSE,
# color = list(
# colorBy = "Surface.Preparation",
# colorScale = htmlwidgets::JS('d3.scale.category10()')
# )
# )
#
# library(GGally)
# library(reshape2)
# #
# #
# ggcorr(mydata) + facet_wrap(~ Contaminate.Type)
# # ggplot(data = melt(cor(mydata[, independent.variables.num])), aes(x=Var1, y=Var2, fill=value)) + geom_tile()
# #
# # Correlation in each subset combining Suraface Prep and Contamination Type
# #
# for (i in levels(mydata$Surface.Preparation)) {
# for (j in levels(mydata$Contaminate.Type)) {
# cat(paste("\nSurface.Preparation =", i, "and Contaminate.Type =", j, ":\n"))
# print(cor(mydata[which(mydata$Surface.Preparation==i & mydata$Contaminate.Type==j), doe.num]))
# # print(ggplot(data = melt(cor(mydata[which(mydata$Surface.Preparation==i & mydata$Contaminate.Type==j), independent.variables.num])), aes(x=Var1, y=Var2, fill=value)) + geom_tile())
# print(ggcorr(mydata[which(mydata$Surface.Preparation==i & mydata$Contaminate.Type==j), doe.num]))
# }
# }